Stroke Order Determination Model for Character Generation

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing AI-based font generation technologies struggle to accurately correct structural information of characters, leading to issues like broken strokes, uneven stroke edges, missing, or redundant strokes, resulting in errors and reduced accuracy in generated characters.

Innovation Solution

A character processing method and apparatus that utilizes a target stroke order determination model combined with spatial and channel attention mechanisms to accurately determine the positions and orders of strokes in characters, thereby reducing errors in character generation.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If style transfer or picture translation technology is used for character generation, then texture correction is improved, but structural information correction deteriorates

Engineering Contradiction:
Improvetexture correction accuracyVSAvoidstructural information accuracy
Core Design Contradiction:
Measurement precisionVSManufacturing precision

Solution Approach 1:

The patent segments the character processing into two independent modules: a stroke structure correction module that extracts and corrects stroke positions and orders, and a style transfer module that handles texture. This segmentation allows each module to specialize in its respective function, preventing the structural information from being lost during style transfer operations.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs stroke structure correction as a preliminary action before style transfer. By first determining the accurate stroke positions and orders, and then applying style transfer to the corrected structure, the system ensures that structural integrity is maintained while still achieving texture correction benefits.

Inventive Principle:
Principle #10Preliminary action

2Productivity

If AI-based font generation is used, then production efficiency is improved, but character accuracy deteriorates

Engineering Contradiction:
Improveproduction efficiencyVSAvoidcharacter accuracy
Core Design Contradiction:
ProductivityVSManufacturing precision

Solution Approach 1:

The patent implements a feedback mechanism where the stroke order determination model predicts stroke sequences, and the system compares these predictions with actual character structures. The model learns from these feedback loops to continuously improve its accuracy in determining stroke positions and orders, thereby reducing errors in generated characters while maintaining high production efficiency.

Inventive Principle:
Principle #23Feedback

3Stability of the object's composition

If conventional stroke generation is used, then character structure is maintained, but stroke breakage and irregularity increase

Engineering Contradiction:
Improvecharacter structure integrityVSAvoidstroke quality
Core Design Contradiction:
Stability of the object's compositionVSManufacturing precision

Solution Approach 1:

The patent changes the parameter representation from pixel-based image data to stroke-based structural parameters. By representing characters in terms of stroke positions, orders, and trajectories, the system can precisely control stroke generation parameters, preventing breakage and irregularity while maintaining structural integrity through mathematically defined stroke paths.

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentUS20250191398A1Character processing method and apparatus, and electronic device and storage medium
Publication Date: 2025.06.12 BEIJING ZITIAO NETWORK TECH CO LTD
  • US20250191398A1 patent drawing
  • US20250191398A1 patent drawing
  • US20250191398A1 patent drawing

AI summary

A character processing method and apparatus, an electronic device and a storage medium. The method includes: acquiring a first image comprising a to-be-processed character; training a target stroke order determination model by combining with a spatial attention mechanism and a channel attention mechanism; and inputting the first image into the target stroke order determination model trained in advance to obtain a target stroke order corresponding to the to-be-processed character.